Echoes from another place - by Jon Treadway

Scholarly Futures

SubscribeSign in

Echoes from another place<br>By asking an LLM to source interesting articles and emerging research relevant to Scholarly Futures, we inadvertently opened up a window into some of the big questions we should all be thinking about.

Jon Treadway<br>Jun 02, 2026

Share

Some time ago, I asked a well known LLM to create a weekly digest that would identify recently published materials relevant to a Substack I was working on. I provided a link to our ‘About page’ to provide some additional context. Sarah and I are fairly well-connected, and we have our ears to the ground, so we tend to have good radars for what is happening. But there has been a lot going on recently, and I was particularly keen to track some of the greyer literature and items outside our peripheral vision that might be worth digesting and bringing to our small but curious audience.<br>So far, so bland.<br>Thanks for reading Scholarly Futures! Subscribe for free to receive new posts and support my work.

Subscribe

Echoes from another place

However, the results were good. Great, actually. The response provided a range of articles that had passed me by, and that were incredibly interesting and absolutely on point for Scholarly Futures. I was a little embarrassed because in many cases they were from sources I regularly read in depth, but that somehow I’d missed.<br>Fantastic either way. The tool had done what I asked, and I marked articles to read for later.<br>Well, you are probably ahead of me here but of course, when I went to read the articles, they were all hallucinations. Not all of them were hallucinations, but the most interesting, the most vibrant sounding ones were.<br>I present here a list of some of the most intriguing pieces that, in the end, I was not able to read, and the short accompanying summary the LLM provided:<br>“China’s New Open Science Push: Infrastructure Before Incentives” — Science|Business

Summary: This article reports on new Chinese investments in national open‑science platforms, including shared data repositories and AI‑ready computing infrastructure. Rather than focusing on researcher mandates, policymakers are prioritising state‑backed infrastructure to enable reuse and scale. The story speaks directly to Open Science and Open Infrastructure, offering a contrasting model to Europe’s more policy‑driven approach. It also raises questions about interoperability and global alignment between Chinese and Western open‑science systems.

Norway proposes new framework for open science infrastructure funding — Research Council of Norway (policy announcement)

Summary: Norway has proposed a new national framework for funding and sustaining open research infrastructure. The policy emphasizes long‑term investment, interoperability, and alignment with European Open Science Cloud initiatives. It also addresses governance challenges and the need for stable institutional support. This is a significant open science policy development, particularly relevant to sustainable infrastructure models.

“Why Open Infrastructure Is a Security Issue, Not Just a Values Choice” — Invest in Open Infrastructure (IOI) Blog

Summary: This independent blog post argues that under‑funded open research infrastructure represents a systemic security risk, from software supply‑chain vulnerabilities to loss of critical community expertise. Drawing on recent incidents, it reframes sustainability as a matter of resilience rather than ideology. The post bridges Open Science & Infrastructure with Research Security, a connection often treated separately in policy debates. It provides language that could resonate with funders and government audiences.

“AI Literacy Is Becoming a Core Research Skill” — Martin Weller, The Ed Techie (Blog)

Summary: In this independent blog post, Weller argues that basic AI literacy is rapidly becoming as essential for researchers as statistical or information literacy. He cautions against superficial training focused only on tools, advocating instead for conceptual understanding of models, limitations, and incentives. This is relevant to AI in Research and Higher Education Reform, especially doctoral training and staff development. The post complements policy‑heavy discussions with a practitioner‑oriented perspective.

US releases updated guidance on AI and dual‑use research risks — Source: National Science and Technology Council (policy report)

This US policy document outlines updated federal guidance on managing dual‑use risks arising from AI‑enabled research. It emphasizes risk assessment frameworks, institutional responsibility, and cross‑agency coordination. Particular attention is given to biosecurity, cybersecurity, and advanced model release practices. The report is a key research security development, with implications for universities, funders, and international collaboration.

I share these hallucinations not to provide a warning that one should use LLMs with...